Measurement of Pancreatic Volume by Abdominal MRI: A Validation Study

Cedars-Sinai Medical Center, Los Angeles, California, United States of America.
PLoS ONE (Impact Factor: 3.23). 02/2013; 8(2):e55991. DOI: 10.1371/journal.pone.0055991
Source: PubMed


To develop abdominal magnetic resonance imaging (MRI) protocol to measure pancreatic volume in humans and to validate it in large animals.
We performed abdominal MRI in eight mini-pigs using a clinical 3T MRI system. We used consecutive parallel abdominal slices, covering the entire pancreas to calculate pancreatic volume. Following MRI, animals were sacrificed, the pancreas was removed, and the volume of the pancreas was measured by water displacement. We used the same MRI protocol to measure pancreatic volume in 21 humans. To assess reproducibility of in vivo measurement we repeated MRI pancreas volume evaluation within 24 hours in additional five humans.
In mini-pigs the measurements of pancreatic volume by MRI and by water displacement were almost identical (R(2) = 0.9867; p<0.0001). In humans the average pancreas volume was 72.7+/-4.5 ml, range from 35.0 to 105.5 ml. This result is in strong agreement with results of previous large postmortem and computed tomography (CT) studies. Repeated measurements of pancreatic volume in humans were highly reproducible. Pancreatic volume measured in vivo was negatively correlated with age, body fat mass, pancreatic TG levels, and visceral fat mass.
These initial results are highly encouraging and our protocol for pancreatic volume estimation in vivo may prove useful in obesity research to follow in vivo changes of pancreatic volume and structure during time course of obesity and type 2 diabetes development.

Download full-text


Available from: Michael D Nelson
  • Source
    • "Though this was a manual process, we believe that PV measurement using MRI planimetry has potential benefits that would outweigh the costs incurred in terms of utilization of manual time and resources. We used a robust method for PV assessment since manual tracings for an irregularly shaped organ like the pancreas is likely still more precise currently than automated techniques [11] [16], particularly on thin contiguous 1 mm MRI sections suitable for clear depiction of pancreas contours. However, with the advent of sophisticated post-processing contouring software , future automated pancreas volume measurements using MRI planimetry may be possible in a fraction of this time as well. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Background/Objectives To compare pancreas volume (PV) measurement using MRI-based planimetry in patients with Type 2 diabetes mellitus (DM) to PV in normoglycemic individuals. Methods Our institutional review board granted approval of this retrospective study with waiver of informed consent. We searched 2296 consecutive abdominal MRI studies performed at our hospital on patients with no pancreas pathology between September 1,2010 and February 28,2013, for those who also had a fasting plasma glucose and/or hemoglobin A1C within six months of the MRI examination. For those patients who met biochemical criteria for DM, we used medication and clinical records to confirm that 32 of these patients had Type 2 DM. The pancreas contours of 32 Type 2 diabetics and 50 normoglycemic individuals were then traced on non-gadolinium T1-weighted 3D fat suppressed gradient echo images by a radiologist trained in abdominal MRI to calculate PV. PV index (PVI) was calculated as PV/weight to adjust PV for each patient’s weight. PVs and PVIs in both cohorts were compared using t-tests and regression models correcting for weight, age and gender. Results Patients with Type 2 DM had significantly lower PVs than normoglycemic individuals (72.7±20.7cm3 versus 89.6±22.7cm3,p<0.001), and significantly lower PVIs (1.0±0.3cm3/kg versus 1.3±0.3cm3/kg,p<0.001). Using regression models, we found that given the same age, weight and gender, the PV in a patient with Type 2 DM was 17.9mL (20%) lower compared to a normoglycemic individual (p<0.001). Conclusion PV is reduced in Type 2 DM compared to normoglycemic individuals and can be measured using MRI without contrast injection.
    Full-text · Article · Jul 2014 · Pancreatology
  • Source
    • "In clinical practice, however, patients with pancreatitis or pancreatic insufficiency routinely undergo cross-sectional imaging to assess for pancreatic changes. Although there are several published protocols in humans to calculate pancreatic volume using CT [8], [9] and MRI [10], [11], [12], [13], [14], methods in animals models are limited to large animals [10]. In this study, we optimized a method for accurately quantifying pancreatic volume in mice using a 7 Tesla micro-MRI and a thin-sliced RARE sequence protocol. "
    [Show abstract] [Hide abstract]
    ABSTRACT: In experimental models of pancreatic growth and recovery, changes in pancreatic size are assessed by euthanizing a large cohort of animals at varying time points and measuring organ mass. However, to ascertain this information in clinical practice, patients with pancreatic disorders routinely undergo non-invasive cross-sectional imaging of the pancreas using magnetic resonance imaging (MRI) or computed tomography (CT). The aim of the current study was to develop a thin-sliced, optimized sequence protocol using a high field MRI to accurately calculate pancreatic volumes in the most common experimental animal, the mouse. Using a 7 Telsa Bruker micro-MRI system, we performed abdominal imaging in whole-fixed mice in three standard planes: axial, sagittal, and coronal. The contour of the pancreas was traced using Vitrea software and then transformed into a 3-dimensional (3D) reconstruction, from which volumetric measurements were calculated. Images were optimized using heart perfusion-fixation, T1 sequence analysis, and 0.2 to 0.4 mm thick slices. As proof of principle, increases in pancreatic volume among mice of different ages correlated tightly with increasing body weight. In summary, this is the first study to measure pancreatic volumes in mice, using a high field 7 Tesla micro-MRI and a thin-sliced, optimized sequence protocol. We anticipate that micro-MRI will improve the ability to non-invasively quantify changes in pancreatic size and will dramatically reduce the number of animals required to serially assess pancreatic growth and recovery.
    Full-text · Article · Mar 2014 · PLoS ONE
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background Impaired glucose tolerance (IGT) and diabetes mellitus (DM) occur more frequently after bone marrow transplantation and total body irradiation (BMT/TBI), but the mechanism is unclear. This study investigates insulin sensitivity, β-cell reserve and pancreatic volume in adult survivors of childhood acute lymphoblastic leukaemia (ALL).Method Survivors (aged 16-26 years) of ALL treated with BMT/TBI (10-14.4Gy) Group 1 (n=20,10M) were compared with a chemotherapy-only Group 2 (n=28,11M). Participants underwent assessments of insulin sensitivity by whole body composite-insulin-sensitivity-index (ISIcomp) from oral glucose tolerance tests (OGTT); first (AIRarg, AIRg, AUCin10) and second (AUC insecond phase) phase insulin responses from arginine-intravenous glucose tolerance tests; and pancreatic volume by abdominal magnetic resonance imaging(MRI). Data were analysed by odds ratio, Chi-square or Fisher's exact tests, Student t-tests, analysis of covariance(ANCOVA) and Pearson's or partial correlations (5% significance).ResultsAbnormal OGTT were documented in Groups 1 (DM=2, IGT=7). Insulin secretion adjusted for insulin sensitivity was lower in Group 1 than Group 2 as a whole [LogAIRarg (p=0.008), logAIRg(p=0.013) and logAUCin10(p=0.014)] and after exclusion of those with abnormal glucose tolerance [logAIRarg(p=0.011), logAIRg(p=0.007) and logAUCin10 (p=0.006)]. Group 1 had lower pancreatic volume than group 2 [52.0(14.2) vs 72.8(23.5), p=0.001] cm3 and results were consistent after adjustment for size by basal surface area (p=0.019). Pancreatic volume correlated with logAIRarg adjusted log ISIcomp (partial correlation =0.34, p=0.025).Conclusions Adult survivors of childhood BMT/TBI for ALL demonstrated reduced β-cell reserve and smaller pancreatic volume; both likely additional aetiological factors, with reduced insulin sensitivity, in their increased risk of diabetes.This article is protected by copyright. All rights reserved.
    No preview · Article · Jul 2014 · Clinical Endocrinology
Show more